Large-Gamut Digital CMYK Exchange Space

CIE Division 1 Leeds, UK

July 4, 2013 William Li, Kodak Mike Rodriguez

Chris Edge, Kodak

1

Overview

• Motivation • Scope • Methodology • Current Status • Evaluation • Next Steps

2

Motivation 1: Predictability

• Consider content created for digital hard-copy output. • Designers create content for print effectively targeted for a

particular printing condition. For example: —CMYK profile for tagging to CMYK image —CMYK profile for print simulation of RGB image (eg, can become

OutputIntent for PDF/X-4) • Targets are effectively chosen in the workflow whether the user

explicitly defines them or not.

• Question: Does target choice affect result?

• Motivation: Reduce choices, but no less than necessary.

3

Motivation 2: Consistency

• Multiple printer types: —A single site may have multiple different types of digital printer. —A single print buyer may send the same file to different sites with

different printers.

• Motivation: Improve consistency of experience between different printers: better for users, better for digital industry.

4

Motivation 3: Blind Exchange

• In case of blind exchange (eg, for PDF/X), not practical to use actual digital printer profile.

• Motivation: Enable designer to create file before knowing exactly which digital printer work will go to, and avoid having to ship printer profile upstream to each designer.

5

Color Workflow from RGB

RGB CMYK

Exchange

CMYK

Digital Press

• Color separate from RGB (AdobeRGB, sRGB, etc.) to CMYK exchange space via perceptual or RC+BPC or…

• Designer adjusts content in light of product of first-stage transform (in CMYK exchange space).

• Re-separate from CMYK exchange to final CMYK (digital press) via RC+BPC with or without channel constraints.

• Channel constraints only required for some devices. Some devices only require black constraint.

Target = SWOP? Fogra39?

Etc? 6

Color Workflow from CMYK

CMYK

Exchange

CMYK

Digital Press

• Design in CMYK exchange space as assigned profile. • Re-separate from CMYK exchange to final CMYK (digital press)

via RC+BPC with or without channel constraints. • Channel constraints only required for some devices. Some

devices only require black constraint.

Target = SWOP? Fogra39?

Etc? 7

Scope

• What does “digital printer” mean?

• For purposes of this work, considering market as main attribute, there are 2 major practical groupings: 1. Document/publication printers (toner, inkjet) aimed at producing

print resembling traditional offset and gravure. 2. Sign printers (mostly inkjet) competing predominantly against

vibrant work produced by flexo, gravure, screen print.

— Constraints on and purpose of substrates, primaries (ink/toner) are different in these markets (eg, UV light-fastness)

— Question: Does the distinction matter?

8

Objective

• Develop and evaluate CMYK exchange space and associated ICC profile suitable for use as design target in creative applications for work targeted at digital printers.

• Resulting CMYK exchange space: —Shall be reasonably proof-able. —Shall not cause undue artifacts.

9

Methodology

• Two types of approach depending on use case: 1. Identical reproduction across all digital printers. 2. Common colour appearance across all digital printers.

• Fitness evaluation:

— Examine gamut size/shape using L* plane slices. — Examine results of running test images through RGB-

>CMYKCMYK color workflow.

• Note: tone reproduction is less critical for digital printers vs. offset press usage, as digital printers are assumed to always use ICC color management.

10

Colorimetric Reproduction

• Implies intersection gamut common to all digital printers.

• Pro: — “guaranteed” colorimetric match among printers – all colours in

gamut

• Con: — Resulting intersection has tiny volume (RPC-3/4) — Approach cuts off competitive distinctiveness of individual

printers. — Use guarantees gamut mapping and/or clipping at first step in

colour workflow (RGBCMYK)

11

Appearance Reproduction

• Implies larger gamut either similar size to average gamut of digital printers, or completely enclosing all digital printer gamuts.

• Pro: —Second-stage transformation consistently mapping down from

similar CMYK shape – all colours observed in design stage can be retained.

—Less possibility of gamut clipping in first-stage transformation.

12

Candidate 1: RPC-7

• Start with ISO 15339-1 RPC-7 (bigger than Fogra39). • RPC-7 is stated as intended for large-gamut and digital devices.

13

RPC-7 vs. RPC-6 & Fogra39

14

RPC-7 RPC-6 Fogra39

RPC-7 vs. AdobeRGB & sRGB

15

RPC-7 AdobeRGB1998 sRGB

L*=75

L*=50

Comparison vs. RPC-7: SC6 TF2

16

RPC-7 Sample-E Sample-F Sample-G Sample-H

Comparison vs. RPC-7: 4 Printers

17

RPC-7 E-Satin KC80 R5-PGls P2-600S

Comparison vs. RPC-7: 4 More Printers

18

RPC-7 F-004 P2-300 PVC NP240

Observations on RPC-7

• Does well enclosing SC6 TF2 (largely toner-based) digital printers (by design).

• Does not do so well with some newer, especially inkjet devices. • At least one document printer has grossly larger gamut than

RPC-7.

• Conclusion: RPC-7 is reasonable, but insufficient.

19

Candidate 2: PRMG Shell

• ICC v4.3 defines shell of Perceptual Reference Medium Gamut. • Design of PRMG is intended to be superset of all graphic arts

output devices. • Colour space created based on PRMG coordinates (primaries,

secondaries, white/dark points). • On review, green secondary moved to align hue angle with ISO

15339-1 RPC. • Profile posted on ICC web site on GASIG Color Experts’ Day

page.

20

Gamut comparisons, Slice through L*=50

PRMG-based profile

PRMG Shell

RPC-7

21

PRMG-MR vs. RPC-7 & RGB

22

RPC-7 PRMG-MR AdobeRGB1998 sRGB

L*=75

L*=50

Comparison vs. PRMG-MR: 4 Printers

23

PRMG-MR F-004 P2-300 PVC NP240

Comparison vs. PRMG-MR: 4 More Printers

24

PRMG-MR F-004 P2-300 PVC NP240

Observations on PRMG-MR

• Mostly encloses all tested digital printer color gamuts. • Possibly too large compared with older SWOP-referenced toner

devices (part of SC6 TF2 set).

• Discussion: —PRMG-MR as the only large-gamut digital printer gamut, or as an

RPC-8 as part of series RPC-1 through 7?

25

Via SWOP to NP240

26

Via RPC-7 to NP240

27

Via PRMG-MR to NP240

28

Image Evaluation

• General observations:

—Converting through US Web Coated SWOP results in somewhat worse results than through the other paths, especially when outputting to larger-gamut digital devices.

29

Next Steps

• Adjustments to PRMG-based exchange space to have: —TR015 neutral scale —Primary TVI curves aligned to 12647-2 Curve A for CMY,

Curve B for K • Consideration for moving magenta and yellow primaries. • Further image evaluations in workflow, and on real-world

printers.

30

Thank you for your attention

Questions?

31